Previous research has demonstrated age-related decline in spatial navigation skills. However, spatial navigation is a complex ability that requires the integration of many component processes that may show differential susceptibility to aging. In particular, successful navigation depends on the accurate computation of distance and direction from self-generated movements through the environment, known as path integration. In the present study, we utilized virtual environments (VE) to investigate age-related differences in participants’ ability to determine linear distances, angular rotations, and angular displacement exclusively from visual input. Results indicated that for shorter lengths, distances can be gauged from visual input alone with relative equivalence between older and younger participants. As distances increased, however, older individuals were less accurate. There were no age differences in estimating angular rotations. Requiring simultaneous integration of distance and rotation (in a triangle completion homing task) resulted in robust age-differences independent of the absolute amount of displacement. For older adults, accuracy in the triangle completion task was highly related to their rotation accuracy while returning to the point of origin, and less so to their distance accuracy. This study demonstrated age-related deficits in the ability to perform visual return-to-origin tasks and suggests one way in which elderly navigation performance may be compromised.